Onset of Preclinical Alzheimer Disease in Monozygotic Twins.

OBJECTIVE: The present work was undertaken to study the genetic contribution to the start of Alzheimer's disease (AD) with amyloid and tau biomarkers in cognitively intact older identical twins. METHODS: We studied in 96 monozygotic twin-pairs relationships between amyloid-beta (Aß) aggregation as measured by the Aß1-42/1-40 ratio in cerebrospinal fluid (CSF; n = 126) and positron emission tomography (PET, n = 194), and CSF markers for Aß production (beta-secretase 1, Aß1-40, and Aß1-38) and CSF tau. Associations among markers were tested with generalized estimating equations including a random effect for twin status, adjusted for age, gender, and apolipoprotein E ε4 genotype. We used twin analyses to determine relative contributions of genetic and/or environmental factors to AD pathophysiological processes. RESULTS: Twenty-seven individuals (14%) had an abnormal amyloid PET, and 14 twin-pairs (15%) showed discordant amyloid PET scans. Within twin-pairs, Aß production markers and total-tau (t-tau) levels strongly correlated (r range = 0.73-0.86, all p < 0.0001), and Aß aggregation markers and 181-phosphorylated-tau (p-tau) levels correlated moderately strongly (r range = 0.50-0.64, all p < 0.0001). Cross-twin cross-trait analysis showed that Aß1-38 in one twin correlated with Aß1-42/1-40 ratios, and t-tau and p-tau levels in their cotwins (r range = -0.28 to 0.58, all p < .007). Within-pair differences in Aß production markers related to differences in tau levels (r range = 0.49-0.61, all p < 0.0001). Twin discordance analyses suggest that Aß production and tau levels show coordinated increases in very early AD. INTERPRETATION: Our results suggest a substantial genetic/shared environmental background contributes to both Aß and tau increases, suggesting that modulation of environmental risk factors may aid in delaying the onset of AD pathophysiological processes. ANN NEUROL 2021;89:987-1000.

Association of amyloid pathology with memory performance and cognitive complaints in cognitively normal older adults: a monozygotic twin study.

Amyloid pathology in cognitively normal older adults has been associated with low memory performance and cognitive complaints, but findings are conflicting. Using a monozygotic twin design, we further explored this relation. We investigated 199 cognitively normal older adults (96 twin pairs) and assessed cognitive performance, cognitive complaints, and amyloid pathology on positron emission tomography and in the cerebrospinal fluid (CSF). Participants were on average 70.5 (SD = 7.6) years and 114 (57%) were female. Amyloid-positron emission tomography abnormality on visual read and lower CSF amyloid-ß 1-42/1-40 ratio were associated with lower Rey visuospatial memory performance (respectively, ß = -0.39 [SE = 0.17], p = 0.02 and ß = 0.15 [SE = 0.07], p = 0.04). Twin analyses showed that CSF amyloid-ß 1-42/1-40 ratio in one twin of a pair could predict visuospatial memory performance in the cotwin (r = 0.20 [SE = 0.10], p = 0.04). Monozygotic twin discordance analyses further showed a probable effect of disease staging on face-name associative memory performance. Our results suggest amyloid aggregation to be associated with lower visuospatial and face-name-associated memory performance in cognitively normal older adults, supporting the view that amyloid pathology leads to memory dysfunction in very early stages of the disease.

The EMIF-AD PreclinAD study: study design and baseline cohort overview.

BACKGROUND: Amyloid pathology is the pathological hallmark in Alzheimer's disease (AD) and can precede clinical dementia by decades. So far it remains unclear how amyloid pathology leads to cognitive impairment and dementia. To design AD prevention trials it is key to include cognitively normal subjects at high risk for amyloid pathology and to find predictors of cognitive decline in these subjects. These goals can be accomplished by targeting twins, with additional benefits to identify genetic and environmental pathways for amyloid pathology, other AD biomarkers, and cognitive decline. METHODS: From December 2014 to October 2017 we enrolled cognitively normal participants aged 60 years and older from the ongoing Manchester and Newcastle Age and Cognitive Performance Research Cohort and the Netherlands Twins Register. In Manchester we included single individuals, and in Amsterdam monozygotic twin pairs. At baseline, participants completed neuropsychological tests and questionnaires, and underwent physical examination, blood sampling, ultrasound of the carotid arteries, structural and resting state functional brain magnetic resonance imaging, and dynamic amyloid positron emission tomography (PET) scanning with [18F]flutemetamol. In addition, the twin cohort underwent lumbar puncture for cerebrospinal fluid collection, buccal cell collection, magnetoencephalography, optical coherence tomography, and retinal imaging. RESULTS: We included 285 participants, who were on average 74.8 ± 9.7 years old, 64% female. Fifty-eight participants (22%) had an abnormal amyloid PET scan. CONCLUSIONS: A rich baseline dataset of cognitively normal elderly individuals has been established to estimate risk factors and biomarkers for amyloid pathology and future cognitive decline.

Aß-oligomer uptake and the resulting inflammatory response in adult human astrocytes are precluded by an anti-Aß single chain variable fragment in combination with an apoE mimetic peptide.

An imbalance between production and clearance of soluble amyloid-ß (Aß) initiates the pathological process in sporadic Alzheimer's disease (AD). Aß-specific antibodies seemed promising as therapeutic option in AD mouse models. In patients, however, vascular side-effects and Aß-antibody complex-induced microglial and/or perivascular macrophage inflammatory responses were encountered. To prevent inflammatory reactions, we designed a single chain variable fragment (scFv-h3D6), based on monoclonal antibody bapineuzumab (mAb-h3D6), but lacking the Fc region. ScFv-h3D6 reduced Aß-oligomer burden and prevented AD-associated behavioral and cellular changes in 3xTg-AD mice. As scFv-h3D6 lacks the Fc-tail, it cannot enhance Fc-receptor mediated Aß clearance by microglia and probably exerts its beneficial effects in 3xTg-AD mice through other mechanisms. ScFv-h3D6 restored the increased apoE and apoJ levels in 3xTg-AD brains back to normal. ApoE and apoJ influence cholesterol transport, Aß aggregation and clearance, and their genetic variants are risk factors for sporadic AD. Astrocytes are constitutive scavengers of soluble Aß from the CNS. We previously found apoE and apoJ to inhibit Aß uptake by adult human astrocytes, in vitro, and thus to potentially protect astrocytes from Aß cytotoxicity. In the present study, scFv-h3D6 and mAb-h3D6 inhibited Aß-oligomer uptake by adult human astrocytes. ApoE- and apoJ- mimetic peptides (MP) affected Aß uptake as well as Aß-induced cytokine release similar to intact apoE and apoJ, without interfering with the strong inhibitory effects of scFv-h3D6 on Aß-oligomer uptake. These results suggest that combining Aß-specific scFv and apoE-MP, that inhibits Aß oligomer-induced cytokine release by astrocytes, could offer advantages over currently used therapeutics.

Unbiased Approach to Counteract Upward Drift in Cerebrospinal Fluid Amyloid-ß 1-42 Analysis Results.

BACKGROUND: Low cerebrospinal fluid (CSF) amyloid-ß 1-42 (Aß 1-42) concentrations indicate amyloid plaque accumulation in the brain, a pathological hallmark of Alzheimer disease (AD). Innotest assay values of Aß 1-42 have gradually increased over the past 2 decades, which might lead to misclassification of AD when a single cutpoint for abnormality is used. We propose an unbiased approach to statistically correct for drift. METHODS: We determined year-specific cutpoints with Gaussian mixture modeling, based on the cross-section of bimodal distributions of Aß 1-42 concentrations in 4397 memory clinic patients. This allowed us to realign year-specific cutpoints as an unbiased method to remove drift from the data. Sensitivity and specificity to detect AD dementia were compared between corrected and uncorrected values. RESULTS: Aß 1-42 values increased 22 pg/mL annually, and this could not be explained by changes in cohort composition. Our approach removed time dependencies [ß (SE) = 0.07 (0.59); P = 0.91]. Statistically correcting for drift improved the sensitivity to detect AD dementia to 0.90 (95% CI, 0.89-0.92) from at least 0.66 (95% CI, 0.64-0.69) based on uncorrected data. Specificity became lower (0.69; 95% CI, 0.67-0.70) vs at most 0.80 (95% CI, 0.79-0.82) for uncorrected data. CONCLUSIONS: This approach may also be useful to standardize Aß 1-42 CSF concentrations across different centers and/or platforms, and to optimize use of CSF biomarker data collected over a long period.

Upward drift in cerebrospinal fluid amyloid ß 42 assay values for more than 10 years.

INTRODUCTION: The best-established cerebrospinal fluid (CSF) biomarkers for Alzheimer's disease are levels of amyloid ß 42 (Aß42), total tau (tau), and phosphorylated tau 181 (ptau). We examined whether a widely used commercial immunoassay for CSF Aß42, tau, and ptau provided stable measurements for more than ∼10 years. METHODS: INNOTEST assay values for CSF Aß42, tau, and ptau from Washington University in St. Louis and VU Medical Center, Amsterdam, were evaluated. RESULTS: Aß42 values as measured by the INNOTEST assay drifted upward by approximately 3% per year over the past decade. Tau values remained relatively stable, whereas results for ptau were mixed. DISCUSSION: Assay drift may reduce statistical power or even confound analyses. The drift in INNOTEST Aß42 values may reduce diagnostic accuracy for Alzheimer's disease in the clinic. We recommend methods to account for assay drift in existing data sets and to reduce assay drift in future studies.

Effects of an Aß-antibody fragment on Aß aggregation and astrocytic uptake are modulated by apolipoprotein E and J mimetic peptides.

Aß-Immunotherapy has long been studied in the treatment of Alzheimer's disease (AD), but not how other molecules involved in the disease can affect antibody performance. We previously designed an antibody fragment, scFv-h3D6, and showed that it precludes Aß-induced cytotoxicity by withdrawing Aß oligomers from the amyloid pathway towards a non-toxic, worm-like pathway. ScFv-h3D6 was effective at the behavioral, cellular, and molecular levels in the 3xTg-AD mouse model. Because scFv-h3D6 treatment restored apolipoprotein E (apoE) and J (apoJ) concentrations to non-pathological values, and Aß internalization by glial cells was found to be decreased in the presence of these apolipoproteins, we now aimed to test the influence of scFv-h3D6 on Aß aggregation and cellular uptake by primary human astrocytes in the presence of therapeutic apoE and apoJ mimetic peptides (MPs). Firstly, we demonstrated by CD and FTIR that the molecules used in this work were well folded. Next, interactions between apoE or apoJ-MP, scFv-h3D6 and Aß were studied by CD. The conformational change induced by the interaction of Aß with apoE-MP was much bigger than the induced with apoJ-MP, in line with the observed formation of protective worm-like fibrils by the scFv-h3D6/Aß complex in the presence of apoJ-MP, but not of apoE-MP. ScFv-h3D6, apoJ-MP, and apoE-MP to a different extent reduced Aß uptake by astrocytes, and apoE-MP partially interfered with the dramatic reduction by scFv-h3D6 while apoJ-MP had no effect on scFv-h3D6 action. As sustained Aß uptake by astrocytes may impair their normal functions, and ultimately neuronal viability, this work shows another beneficence of scFv-h3D6 treatment, which is not further improved by the use of apoE or apoJ mimetic peptides.

Clusterin Levels in Plasma Predict Cognitive Decline and Progression to Alzheimer's Disease.

BACKGROUND: Increased clusterin levels have been reported in brain, cerebrospinal fluid (CSF), and plasma of Alzheimer's disease (AD) patients. Because changes are also observed in mild cognitive impairment (MCI), a possible relationship between clusterin levels and early neurodegenerative changes in AD was suggested. OBJECTIVES: To determine whether clusterin concentrations could 1) serve as a diagnostic marker for AD, 2) predict disease progression in MCI, and 3) correlate with AD-biomarkers. METHODS: Clusterin levels in CSF and plasma, as well as AD biomarker levels of Aß42, Tau, and pTau in CSF and Mini-Mental State Examination scores (MMSE) were determined in 67 controls, 50 MCI, and 107 AD patients. Repeated MMSE was obtained for 44 MCI and 72 AD patients after, on average, 2.7 years. RESULTS: Elevated clusterin concentrations in plasma, but not in CSF, were a risk factor for AD (HR 18.6; 95% CI 2.8-122), and related to cognitive decline in MCI (r =-0.38; p <  0.01). An inverse relation between plasma clusterin levels and cognitive decline was observed in AD patients (r = 0.23; p≤0.05). In CSF, but not in plasma, clusterin levels correlated with Tau and pTau in all groups. CONCLUSION: Elevated plasma clusterin levels in MCI confer an increased risk for progression to AD, and more rapid cognitive decline. We speculate that clusterin levels in CSF may reflect its involvement in the earliest neurodegenerative processes associated with AD pathology. Whereas neither clusterin levels in CSF nor in plasma had diagnostic value, plasma clusterin levels may serve as a prognostic marker for AD.

Persistent reduction of hippocampal glutamine synthetase expression after status epilepticus in immature rats.

Mesiotemporal sclerosis (MTS), the most frequent form of drug-resistant temporal lobe epilepsy, often develops after an initial precipitating injury affecting the immature brain. To analyse early processes in epileptogenesis we used the juvenile pilocarpine model to study status epilepticus (SE)-induced changes in expression of key components in the glutamate-glutamine cycle, known to be affected in MTS patients. SE was induced by Li(+) /pilocarpine injection in 21-day-old rats. At 2-19 weeks after SE hippocampal protein expression was analysed by immunohistochemistry and neuron damage by FluoroJade staining. Spontaneous seizures occurred in at least 44% of animals 15-18 weeks after SE. As expected in this model, we did not observe loss of principal hippocampal neurons. Neuron damage was most pronounced in the hilus, where we also detected progressive loss of parvalbumin-positive GABAergic interneurons. Hilar neuron loss (or end-folium sclerosis), a common feature in patients with MTS, was accompanied by a progressively decreased glutamine synthetase (GS)-immunoreactivity from 2 (-15%) to 19 weeks (-33.5%) after SE. Immunoreactivity for excitatory amino-acid transporters, vesicular glutamate transporter 1 and glial fibrillary acidic protein was unaffected. Our data show that SE elicited in 21-day-old rats induces a progressive reduction in hilar GS expression without affecting other key components of the glutamate-glutamine cycle. Reduced expression of glial enzyme GS was first detected 2 weeks after SE, and thus clearly before spontaneous recurrent seizures occurred. These results support the hypothesis that reduced GS expression is an early event in the development of hippocampal sclerosis in MTS patients and emphasize the importance of astrocytes in early epileptogenesis.

Apolipoproteins E and J interfere with amyloid-beta uptake by primary human astrocytes and microglia in vitro.

Defective clearance of the amyloid-ß peptide (Aß) from the brain is considered a strong promoter in Alzheimer's disease (AD) pathogenesis. Astrocytes and microglia are important mediators of Aß clearance and Aß aggregation state and the presence of amyloid associated proteins (AAPs), such as Apolipoproteins E and J (ApoE and ApoJ), may influence Aß clearance by these cells. Here we set out to investigate whether astrocytes and microglia differ in uptake efficiency of Aß oligomers (Aßoligo ) and Aß fibrils (Aßfib ), and whether the Aß aggregation state and/or presence of AAPs affect Aß uptake in these cells in vitro. Adult human primary microglia and astrocytes, isolated from short delay post-mortem brain tissue, were exposed to either Aßoligo or Aßfib alone or combined with a panel of certain AAPs whereafter Aß-positive cells were quantified using flow cytometry. Upon exposure to Aß combined with ApoE, ApoJ, α1-antichymotrypsin (ACT) and a combination of serum amyloid P and complement C1q (SAP-C1q), a clear reduction in astrocytic but not microglial Aßoligo uptake, was observed. In contrast, Aßfib uptake was strongly reduced in the presence of AAPs in microglia, but not in astrocytes. These data provide the first evidence of distinct roles of microglia and astrocytes in Aß clearance. More importantly we show that Aß clearance by glial cells is negatively affected by AAPs like ApoE and ApoJ. Thus, targeting the association of Aß with AAPs, such as ApoE and ApoJ, could serve as a therapeutic strategy to increase Aß clearance by glial cells.

The effect of amyloid associated proteins on the expression of genes involved in amyloid-ß clearance by adult human astrocytes.

Astrocytes appear to be important mediators in the clearance of amyloid beta1-42 (Aß), the key component of senile plaques characteristic of Alzheimer's disease (AD). Recently, we found the amyloid associated proteins (AAPs) α1-antichymotrypsin (ACT), apolipoprotein J and E (ApoJ and ApoE) and a mixture of serum amyloid P (SAP) and C1q (SAP-C1q) to modify Aß-uptake by human astrocytes. Here we investigated the effect of oligomeric (Aßoligo) and fibrillar Aß (Aßfib), alone and in combination with a panel of AAPs on the astrocytic expression of genes proposed to be involved in Aß-uptake and degradation. Primary human astrocytes (isolated from non-demented control (n=4) and AD patient (n=4) brain specimens) were exposed to either Aßoligo or Aßfib preparations with or without the above mentioned AAPs. Quantitative gene expression analysis of Aß-receptors Scavenger receptor B1 (SCARB1), macrophage receptor with collagenous structure (MARCO) and low density lipoprotein receptor related protein-2 (LRP2 or megalin) as well as of Aß-degrading enzymes neprilysin (NEP), insulin-degrading enzyme (IDE) and metalloproteinase-9 (MMP-9) was performed by real-time PCR. Basal expression of NEP, IDE and SCARB1 was easily detected whereas expression of MARCO, LRP2 and MMP-9 could only be detected upon pre-amplification. Basal expression of NEP, IDE and SCARB1 did not change upon exposure to Aßoligo or Aßfib alone in any of the investigated astrocyte cultures. Interestingly NEP expression was increased upon exposure to ApoE in combination with both Aß-preparations, and also SCARB1 expression was induced upon treatment with ApoE in combination with Aßfib in astrocytes from non-demented controls. Further, SAP-C1q increased SCARB1 expression in control astrocytes when combined with Aßoligo. These alterations were not found in astrocytes from AD patients. Thus, we conclude that Aß alone apparently does not affect the astrocytic expression of IDE, NEP or SCARB1. However, NEP and SCARB1 expression is increased in astrocytes from non-demented subjects when exposed to Aß combined with AAPs like ApoE. These astrocytic gene expression-regulatory mechanisms appear to be defective in AD and thus might contribute to the development and progression of AD pathology.

Evaluation of intrathecal serum amyloid P (SAP) and C-reactive protein (CRP) synthesis in Alzheimer's disease with the use of index values.

Serum amyloid P (SAP) and C-reactive protein (CRP) are proteins involved in innate immunity. The expression of SAP and CRP is increased in Alzheimer's disease (AD) brain tissue, compared to healthy controls. Although both proteins are found in cerebrospinal fluid (CSF), their origin is unclear. We investigated if increased local production of SAP and CRP in AD brain results in higher levels in CSF with the use of index values. To study this, SAP, CRP, and albumin levels were determined in CSF and serum samples of 30 control (65 ± 11 years; 57% female) and 140 AD subjects (65 ± 9 years; 53% female). To correct for inter-individual differences in protein diffusion from blood to CSF, quotients (Q =CSF/serum) of SAP, CRP, and albumin and index values (Qprotein/Qalb) were calculated. The results showed no significant differences in SAP and CRP index values between control and AD subjects, although eight percent of individual AD patients showed evidence of intrathecal SAP or CRP production using the Reiber hyperbolic model. Interestingly, the SAP index value was much lower than expected, based on its molecular size. In conclusion, these data suggest that local production of SAP and CRP in the AD brain does not substantially contribute to the CSF levels.

Astrocytic A beta 1-42 uptake is determined by A beta-aggregation state and the presence of amyloid-associated proteins.

Intracerebral accumulation of amyloid-beta (A beta) leading to A beta plaque formation, is the main hallmark of Alzheimer's disease and might be caused by defective A beta-clearance. We previously found primary human astrocytes and microglia able to bind and ingest A beta 1-42 in vitro, which appeared to be limited by A beta 1-42 fibril formation. We now confirm that astrocytic A beta-uptake depends on size and/or composition of A beta-aggregates as astrocytes preferably take up oligomeric A beta over fibrillar A beta. Upon exposure to either fluorescence-labelled A beta 1-42 oligomers (A beta(oligo)) or fibrils (A beta(fib)), a larger (3.7 times more) proportion of astrocytes ingested oligomers compared to fibrils, as determined by flow cytometry. A beta-internalization was verified using confocal microscopy and live-cell imaging. Neither uptake of A beta(oligo) nor A beta(fib), triggered proinflammatory activation of the astrocytes, as judged by quantification of interleukin-6 and monocyte-chemoattractant protein-1 release. Amyloid-associated proteins, including alpha1-antichymotrypsin (ACT), serum amyloid P component (SAP), C1q and apolipoproteins E (ApoE) and J (ApoJ) were earlier found to influence A beta-aggregation. Here, astrocytic uptake of A beta(fib) increased when added to the cells in combination with SAP and C1q (SAP/C1q), but was unchanged in the presence of ApoE, ApoJ and ACT. Interestingly, ApoJ and ApoE dramatically reduced the number of A beta(oligo)-positive astrocytes, whereas SAP/C1q slightly reduced A beta(oligo) uptake. Thus, amyloid-associated proteins, especially ApoJ and ApoE, can alter A beta-uptake in vitro and hence may influence A beta clearance and plaque formation in vivo.

BACE1 activity in cerebrospinal fluid and its relation to markers of AD pathology.

Several studies have shown that reduced amyloid-beta 1-42 (Abeta(42)) and increased tau levels in cerebrospinal fluid (CSF) reflect increased Alzheimer's disease (AD) pathology in the brain. beta-site APP cleaving enzyme (BACE1) is thought to be the major beta-secretase involved in Abeta production in the brain, and therefore we investigated the relation between BACE1 activity and CSF markers Abeta(40), Abeta(42), total tau (t-tau), and tau phosphorylated at threonine 181 (p-tau) in CSF of control (n=12), mild cognitive impairment (n=18), and AD (n=17) subjects. Patients were classified according to their Abeta(42), t-tau, and p-tau CSF biomarker levels, with either an AD-like biomarker profile (two or three biomarkers abnormal: Abeta(42) < 495 pg/ml in combination with t-tau > 356 pg/ml, and/or p-tau > 54 pg/ml) or a normal biomarker profile (

CSF levels of PSA and PSA-ACT complexes in Alzheimer's disease.

BACKGROUND: Prostate-specific antigen (PSA) is a serine protease that in serum, is predominantly found complexed to the serine protease inhibitor alpha1-antichymotrypsin (ACT). ACT co-localizes with amyloid plaques in Alzheimer's disease (AD) brain and both PSA and ACT are detectable in cerebrospinal fluid (CSF). Therefore, we aimed to determine whether PSA is produced in the brain and whether PSA and PSA-ACT complex levels in CSF can be used as a biomarker for AD. METHODS: Levels of ACT and PSA-ACT were determined by sandwich enzyme-linked immunosorbent assay in CSF and serum samples of AD (n = 16), frontotemporal lobe dementia (FTLD) (n = 19), mild cognitively impaired (MCI) patients (n = 19) and controls (n = 12). Total PSA was determined in a non-competitive immunoassay. Reverse transcriptase-polymerase chain reaction (RT-PCR) for PSA was performed on postmortem hippocampus and temporal cortex specimens from control and AD cases. RESULTS: PSA is expressed in the brain, as detected by RT-PCR. PSA and PSA-ACT complexes were detectable in CSF of almost all male and only very few female subjects. The levels of PSA and PSA-ACT complexes in CSF did not differ between AD, FTLD, MCI and control groups. PSA CSF/serum quotients highly correlated with albumin CSF/serum quotients. Furthermore, the hydrodynamic radius of PSA was found to be 3 nm and the theoretical PSA quotient, derived from the Felgenhauer plot, corresponded well with the measured PSA quotient. CONCLUSIONS: PSA is locally produced in the human brain; however, brain PSA hardly contributes to the CSF levels of PSA. PSA and PSA-ACT levels in CSF are not suitable as a biomarker for AD.

Serum amyloid p component as a biomarker in mild cognitive impairment and Alzheimer's disease.

BACKGROUND: Serum amyloid P component (SAP), present in amyloid-beta (Abeta) plaques in Alzheimer's disease (AD), may protect Abeta deposits against proteolysis, thereby promoting plaque formation. The aim was to investigate if SAP levels in cerebrospinal fluid (CSF) and serum can be used to discriminate controls, AD and mild cognitive impairment (MCI) patients, and to identify incipient AD among MCI patients. METHODS: SAP levels in CSF and serum were determined in 30 controls, 67 MCI and 144 AD patients. At follow-up, 39 MCI patients had progressed to dementia, while 25 had remained stable (mean follow-up time: 2.6 +/- 1.0 and 2.1 +/- 0.8 years). RESULTS: Cross-sectionally no differences were found in SAP levels in CSF and serum between the groups. MCI patients that had progressed to dementia at follow-up had lower CSF SAP levels (13 microgram/l, range 3.3-199.3 microgram/l) than MCI nonprogressors (20.2 microgram/l, range 7.0-127.7 microgram/l; p < 0.05) [corrected]. A low CSF SAP level was associated with a 2-fold increased risk of progression to AD (hazard ratio = 2.2; 95% confidence interval = 0.9-5.4). CONCLUSION: Our data suggest that measurement of CSF SAP levels can aid in the identification of incipient AD among MCI patients.